Harness, the Autonomous SDLC Platform company, announced on October 8, 2026 that it has acquired select Augment Code assets, including Cosmos, the Auggie CLI, and the Code Context Engine. The team behind these products will join Harness to accelerate its work toward an increasingly autonomous software development lifecycle. The deal introduces the Harness Cosmos Software Factory Agent, connecting code and delivery data from idea to production.
Acquisition Brings Code and Delivery Context Together
Under the agreement, Cosmos becomes the Harness Cosmos Software Factory Agent focused on automating engineering work from idea to code. Augment's Code Context Engine will combine with Harness's Software Delivery Knowledge Graph to connect code-level insight with delivery and production data. Once Cosmos produces a reviewed pull request, Harness delivery, security, runtime, and cost agents carry the change into production.
How the Harness Cosmos Agent Works
Harness Cosmos operates as a fleet of agents triggered by requirements, tickets, or bugs. Each agent plans the change, writes code and tests, and opens a pull request in an isolated virtual machine. Engineers are pulled in only where judgment is needed, such as design approval or final merge, while agents run continuous loops across ticket, code, and review.
The Code Context Engine maintains a live map of the codebase, and model routing matches each task to the right model. Teams can fork prebuilt Experts such as Project Builder, PR Author, Deep Reviewer, and PR Fixer, then tune each one to their codebase. Integrations with GitHub, Jira, and Slack, along with versioning and budget controls, let teams run autonomous coding across the organization.
Shared Context from Code to Production
Before a change is written, Harness Cosmos pulls together delivery context such as test cases, security requirements, past incidents, and remediations. Downstream Harness agents use codebase context to assess dependencies, risk, and required validation. If something fails during testing, security review, or deployment, the agents can then loop until resolution.
Leadership Perspective
"AI's impact on software will not be measured by how much code it generates, but by how much valuable software reaches customers," said Harness founder and CEO Jyoti Bansal. He explained that Augment Code built Cosmos to create a software factory for repeatable engineering work, and the acquisition unites code, delivery, and production context. Bansal added that when downstream Harness agents identify an issue while testing, securing, or deploying code, they send it back to Cosmos for correction.
"We set out to bring AI to the entire software development lifecycle, starting with context-aware coding," said Igor Ostrovsky, co-founder and CTO of Augment Code. He described joining Harness as the next step toward that vision, noting that engineering work can become increasingly autonomous with engineers stepping in only where judgment really matters. Ostrovsky said software factories have transformed how companies build software, and this acquisition makes the same transformation available to every team on the Harness platform.
Availability and Company Momentum
Harness Cosmos is now available, and the company is working to connect its workflows with the policies, permissions, and approvals that already govern the Harness platform. The broader Harness platform includes the Software Delivery Agent, Security Testing Agent, Runtime Protection Agent, and Cost Management Agent, tied together by the Agent Harness. United Airlines, Morningstar, and Choice Hotels use Harness to accelerate releases by up to 75%, cut cloud costs by up to 60%, and improve efficiency up to 10x.
The acquisition positions Harness to unify agent-driven development with production operations through shared code and delivery context. By combining Augment's codebase understanding with Harness's delivery knowledge, the company aims to make more engineering work autonomous while preserving human judgment at critical points. The move signals continued expansion of the autonomous SDLC category and deeper integration between coding tools and delivery platforms.